Spatial and temporal patterns in water quality were analysed in
the multivariate domain using Principal Component Analysis (PCA)
on untransformed, standardised water quality variables complemented by Analysis of Variance (ANOVA)
for univariate measures (Underwood, 1981). The ANOVA models
were factorial, comprising three terms: (a) bay, (b) date, and (c)
depth. Following significant main or interaction effects, post hoc
comparisons were performed with Student Newman Keuls (SNK)
tests.